From:"Bonate, Peter"
Subject:[NMusers] Wrong model
Date:Fri, 12 Jul 2002 11:07:15 -0500
Hi, everyone. Its generally recognized that with sparse data from a Phase III study we often need
to fit a simpler compartmental model than the one that could be obtained from rich Phase I data,
e.g., 1-compartment with sparse data but 2-compartment with rich data. Can someone point me to
some references that examine what impact this simplification has on the estimates of the common pk
parameters. For example, how biased is clearance or central volume. Thanks alot.
pete
Peter L. Bonate, PhD
Director, Pharmacokinetics
ILEX Oncology, Inc
4545 Horizon Hill Blvd
San Antonio, TX 78229
phone: 210-949-8662
fax: 210-949-8487
email: pbonate@ilexonc.com
Wrong model
From:"HUTMACHER, MATTHEW [Non-Pharmacia/1820]"
Subject:RE: [NMusers] Wrong model
Date:Fri, 12 Jul 2002 12:57:06 -0400
Hello Peter,
Often CL is not badly biased, with volume being biased and estimated
somewhere between Vc and Vss. Ka is often horribly estimated but its
pathology does affect the estimation of CL (in my experience).
See Kowalski and Hutmacher. Design evaluation for a population PK study
using clinical trial simulations. Stats in Medicine 2001; 20:75-91
Hope this helps.
Matt
From: "Sambol, Nancy"
Subject:Re: [NMusers] Wrong model
Date:Fri, 12 Jul 2002 10:52:24 -0700
Hi Peter and Nonmem users,
The impact of structural model misspecification on other parts of the model was
addressed in the following paper. A bit of a surprise for us (at least then) was that
the impact of the misspecification also depends on your statistical model
(whether parameters are allowed to covary or not).
Wade JR, Beal SL and Sambol NC. Interaction between structural, statistical
and covariate models in population pharmacokinetic analysis.
J Pharmacokin Biopharm 1994; 22: 165-177
Hope this helps.
Nancy
Nancy Sambol, Pharm.D.
Associate Clinical Professor
Director of Data Analysis Group